Attachment of a hydrogen-bonding carboxylate side chain to an [FeFe]- hydrogenase model complex: Influence on the catalytic mechanism

Article abstract of DOI:10.1002/chem.200902278

Azapropanedithiolate (adt)- bridged model complexes of [FeFe]-hydrogenase bearing a carboxylic acid functionality have been designed with the aim of decreasing the potential for reduction of protons to hydrogen. Protonation of the bisphosphine complexes 4-6 has been studied by in situ IR and NMR spectroscopy, which revealed that protonation with triflic acid most likely takes place first at the N-bridge for complex 4 but at the Fe-Fe bondfor complexes 5 and 6. Using an excess of acid, the diprotonated species could also be observed, but none of the protonated species was sufficiently stable to be isolated in a pure state. Electrochemical studies have provided an insight into the catalytic mechanismsunder strongly acidic conditions, and have also shown that complexes 3 and 6 are electro-active in aqueous solution even in the absence of acid, presumably due to hydrogen bonding. Hydrogen evolution, driven by visible light, has been observed for three-component systems consisting of [Ru(bpy)₃]²⁺ ,complex 1, 2, or 3, and ascorbic acid in CH₃CN/D₂O solution by on-line mass spectrometry.

Full text of DOI:10.1002/chem.200902278

FULL PAPER
DOI: 10.1002/chem.200902278
Attachment of a Hydrogen-Bonding Carboxylate Side Chain to an [FeFe]-
Hydrogenase Model Complex: Influence on the Catalytic Mechanism
Weiming Gao,[a] Junliang Sun,[b] Torbjçrn ꢀkermark,[c] Mingrun Li,[b] Lars Eriksson,[b]
Licheng Sun,[d] and Bjçrn ꢀkermark*[a]
Abstract: Azapropanedithiolate (adt)-
bridged model complexes of [FeFe]-hy-
drogenase bearing a carboxylic acid
functionality have been designed with
the aim of decreasing the potential for
reduction of protons to hydrogen. Pro-
tonation of the bisphosphine complexes
4–6 has been studied by in situ IR and
NMR spectroscopy, which revealed
that protonation with triflic acid most
likely takes place first at the N-bridge
for complexes 5 and 6. Using an excess
of acid, the diprotonated species could
also be observed, but none of the pro-
tonated species was sufficiently stable
to be isolated in a pure state. Electro-
chemical studies have provided an in-
sight into the catalytic mechanisms
under strongly acidic conditions, and
have also shown that complexes 3 and
6 are electro-active in aqueous solution
even in the absence of acid, presuma-
bly due to hydrogen bonding. Hydro-
gen evolution, driven by visible light,
has been observed for three-compo-
nent systems consisting of [RuACHTUNGTRENNUNG
(bpy)3]2+
, complex 1, 2, or 3, and ascorbic acid
in CH3CN/D2O solution by on-line
mass spectrometry.
Keywords: electrochemistry · elec-
tron transfer · enzyme model · iron-
hydrogenase · proton reduction
À
for complex 4 but at the Fe Fe bond
Introduction
even with these advanced structures, protons are reduced to
hydrogen at a much more negative potential, usually <
À1.5 V versus Fc/Fc+ (À1.12 V vs. SCE; À0.88 V vs. NHE)
than with the natural hydrogenases (ca. À0.4 V vs. NHE;
À1.02 V vs. Fc/Fc+ at pH 7).[1] One explanation is that,
unlike the hydrogenases, the synthetic models do not usually
contain ligands such as cyanide; another could be that the
ligand field of the natural protein surrounding is very differ-
ent from those in the synthetic models.
During studies of synthetic models, it has often been
found that protonation with strong acids is necessary to ach-
ieve hydrogen production even at potentials of about
À1.5 V versus Fc/Fc+. The reason is that either protonation
of the nitrogen of the adt bridge or the formation of iron hy-
drides, or indeed both, is required for proton reduction at a
reasonable potential.[2] In adt-bridged model complexes, on
the other hand, the interaction between the NH proton and
the FeH could aid in producing hydrogen at a reasonable
potential.[2] In nature, however, strong acids clearly cannot
play an important role. It therefore occurred to us that in
the natural hydrogenases nearby carboxyl groups could play
a role, both through hydrogen bonding and by supplying
protons, allowing the formation of hydrides under mild con-
ditions. By participating in proton-coupled electron transfer
(PCET) from the required electron donors, carboxyl groups
should also be capable of lowering the redox potentials re-
Efforts to determine the structure of the active site of
[FeFe]-hydrogenases[1] have led to extensive studies of
models, such as azapropanedithiolate (adt) complexes of the
type [Fe2(CO)6ÀxLxACHUNTGREN{NUG (SCH2)2NR}] (L=PMe3). However,
[a] Dr. W. Gao, Prof. B. ꢀkermark
Department of Organic Chemistry, Arrhenius Laboratory
Stockholm University, 10691 Stockholm (Sweden)
Fax : (+46)8-154908
E-mail: bjorn.akermark@organ.su.se
[b] Dr. J. Sun, Dr. M. Li, Dr. L. Eriksson
Division of Structural Chemistry, Arrhenius Laboratory
Stockholm University, 10691 Stockholm (Sweden)
[c] Dr. T. ꢀkermark
Applied Electrochemistry
School of Chemical Science and Engineering
Royal Institute of Technology (KTH), 10044 Stockholm (Sweden)
[d] Prof. L. Sun
Department of Chemistry
Royal Institute of Technology (KTH)
10044 Stockholm (Sweden)
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.200902278: Crystal structures of
dimers of 1–3; mechanisms A and B for CECE; cyclic voltammo-
grams of 1–6 in the presence of incremental amounts of HOTf; and
isotope distributions of evolved hydrogen with 2 and 3.
Chem. Eur. J. 2010, 16, 2537 – 2546
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2537

quired for the generation of hydrogen.[3] Although Darens-
bourg et al. and Talarmin et al. have reported a few Fe2S2
complexes bearing carboxylate groups that were shown to
be efficient electrocatalysts in the presence of weak acids,
strongly negative potentials were still required.[4] We there-
fore decided to prepare some complexes of the type
Table 1. IR data for complexes 1–6 as well as two pdt-diiron and two
adt-diiron complexes for comparison.
Complex n(CO)Fe [cmÀ1
]
n
(C=O)OH [cmÀ1
]
Ref.
1
2
3
4
5
6
2073(m), 2029(s), 1991(s)
2073(m), 2029(s), 1991(s)
2074(m), 2033(s), 1994(s)
1977(m), 1939(s), 1899(s)
1979(m), 1938(s), 1896(s)
1978(m), 1941(s), 1898(s)
2077(m), 2034(s), 1992(s), 1978(sh)
1984(m), 1948(vs), 1903(s)
1981(m), 1943(vs), 1907(s), 1892(sh)
1982(m), 1945(vs), 1901(s)
1691
1709
1710
1694
1665
1659
1730
1729
—
[a]
[a]
[a]
[a]
[a]
[a]
[4a]
[4a]
[2a]
[6a]
[Fe2(CO)6ÀxLxACHTUNGTRENNUNG{(SCH2)2NR}] (L=PMe3) containing a carbox-
ylate group capable of interacting with the adt nitrogen, and
to study the redox potentials for proton reduction catalyzed
by these complexes.
7[b]
8[b]
9[c]
10[c]
—
[a] This work. [b] Spectrum recorded in THF. [c] Spectrum recorded in
CH3CN.
Results and Discussion
Preparation of complexes 1–6: The carboxylate-substituted
diiron adt complexes 1–3 were synthesized according to
Rauchfussꢂs route[5] by reacting [Fe2(CO)
(HOCH2S)2] with
Properties of the complexes: When two of the six carbonyls
were replaced by PMe3 ligands, the average position of the
three n(CO)Fe bands in the IR spectra shifted by about
90 cmÀ1 to higher frequencies. The 31P NMR shifts of the
complexes are similar, d=22.3 ppm for complexes 4 and 5
and d=22.7 ppm for 6, and fall in the same range as the
published shifts for related complexes.[2a,4a,6a,b]
It is interesting to compare
the corresponding amino acids at À788C and then warming
to room temperature. This gave the complexes 1–3 in mod-
erate isolated yields (30–40%). Complexes 4–6 bearing two
PMe3 groups were then obtained in respective yields of
around 50% by reacting complexes 1–3 with four equiva-
lents of PMe3 in refluxing THF. The n(CO) bands in the IR
the IR frequencies of the car-
boxyl groups of the complexes.
Among the hexacarbonyl com-
plexes,
a low carbonyl shift
(1691 cmÀ1), indicative of strong
hydrogen bonding, was ob-
served for the carboxyl group
of complex 1. This is presuma-
bly due to the formation of hy-
drogen-bonded
dimers,
as
shown by the crystal structure
determination. In contrast, the
carbonyl frequencies for 2 and
3 are around 1710 cmÀ1, indicat-
ing only weak hydrogen bond-
ing, despite the fact that these
complexes also form dimers in
the crystalline state, with simi-
À
lar O H···O distances (2.644 ꢀ
for 1, 2.642 ꢀ for 2, and
2.669 ꢀ for 3; see the Support-
ing Information). This may pos-
sibly be because the carboxyl
group is conjugated with the ar-
omatic system in 1.
The replacement of two car-
spectra of complexes 1–6, along with those of previously re-
bonyl ligands by PMe3 induced essentially no change on
going from complex 1 to 4, which seems to retain its dimeric
structure, but for complexes 5 and 6 the low carbonyl fre-
quencies of 1665 and 1659 cmÀ1 indicate strong hydrogen
bonding. A reasonable explanation is that the increased
electron density on the diiron cluster that results from the
presence of the phosphine donor ligands also renders the ni-
ported
[Fe2(CO)6{m-
(PMe3)2{m-(SCH2)2CHCOOH}] (8), and adt complexes
[Fe2(CO)4A(PMe3)2A{(m-SCH2)2NCH2C6H5}] (9) and [Fe2(CO)4-
(PMe3)2A{(m-SCH2)2NCH2C6H4-2-Br}] (10), are summarized
in Table 1.[2a,4a,6a]
propanedithiolate
(pdt)
complexes
ACHTUNGTRENNUNG(SCH2)2CHCOOH}] (7) and [Fe2(CO)4-
A
ACHTUNGTRENNUNG
C
CHTUNGTRENNUNG
A
CHTUNGTRENNUNG
2538
www.chemeurj.org
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 2537 – 2546

ACHTUNGTRENNUNG[FeFe]-Hydrogenase Model Complexes with Carboxylate Side-Chains
FULL PAPER
trogen sufficiently basic to participate in intramolecular hy-
drogen bonding to the carboxyl groups. This is shown by the
fact that the IR frequency (1660 cmÀ1) is unchanged when
the spectrum of complex 6 is measured from a sample in
THF, which should break intermolecular hydrogen bonds
but not affect reasonably strong intramolecular bonds.[7] In
contrast, the IR frequency of the carboxyl groups in com-
plex 8, which lacks the nitrogen (1729 cmÀ1 in THF solu-
tion), shows that no intramolecular bond is formed in this
complex.
It has previously been proposed that the protonated nitro-
gen of the adt bridge can participate by hydrogen bonding
to an initially formed iron hydride and thereby serve to pro-
tonate the hydride to give molecular hydrogen.[2b,c] Presuma-
bly, hydrogen bonding to the carboxylate can help to in-
crease the mobility of the proton attached to the nitrogen.
In addition, such hydrogen bonding could decrease the
redox potential for the initial formation of iron hydride by
facilitating proton-coupled electron transfer. This is illustrat-
ed for complex 6 in Scheme 1. First, proton-coupled reduc-
Molecular structures: The crystal structures of complexes 1–
3 are shown in Figure 1, together with the atom-labeling
system. Selected bond lengths and angles are given in
Table 2.
Table 2. Selected metric data (bond lengths [ꢀ] and angles [8]) for com-
plexes 1–3.
1
2
3
Fe(1)-Fe(2)
Fe(1)-S(1)
Fe(1)-S(2)
2.506(2)
2.255(3)
2.250(3)
67.64(9)
84.85(11)
4.443
2.500(3)
2.254(5)
2.252(5)
67.53(13)
85.03(16)
2.798
2.5029(9)
2.2523(14)
2.2545(14)
67.59(4)
84.24(5)
2.486
Fe(1)-S(1)-Fe(2)
S(1)-Fe(1)-S(2)
C···N[a]
O(H)···N[b]
5.709
3.088
3.598
[a] The distance between the carbon (of the carboxylic acid) and the
bridge nitrogen atom. [b] The nonbonding distance.
Each Fe2S2 subunit has a square-pyramidal structure. The
À
Fe Fe distances in complexes 1 (2.506 ꢀ), 2 (2.500 ꢀ), and
3 (2.503 ꢀ) fall within the range commonly found for diiron
model complexes.[6,8,9] The structure of 1 is quite similar to
that reported for the related complex [Fe2(CO)6ACHTUNGTRENNUNG{(m-
SCH2)2NCH2C6H4-2-Br}]. The sum of the angles at the
bridge N atom of 1 is 3558, which is very similar to that in
[Fe2(CO)6ACHTNURTGNEUNG{(m-SCH2)2NCH2C6H4-2-Br}] (3588), suggesting
near sp2-hybridization. The corresponding sum of the angles
at N in 2 is similar (3558), while it is 3408 in 3, roughly con-
sistent with sp3-hybridization at the nitrogen. This difference
between 2 and 3 is surprising, since they both form carboxyl-
ic dimers in the solid state, with similar IR carbonyl fre-
À
quencies (Table 1). However, in the solid state, the O H···O
bond is distinctly longer in 3 and the carboxyl group is ori-
ented in such a way that it may also be involved in a hydro-
gen bond to the adt nitrogen, whereas the orientation in 2
makes this difficult.
Protonation of the complexes: We first studied the protona-
tion of complexes 1–6 by a strong acid (HOTf, pKa =2.60).
For complexes 1–3, protonation was found to occur only at
the adt nitrogen, even in the presence of five equivalents of
HOTf, with a characteristic blue shift of the carbonyl fre-
quencies (Dn(CO)Fe ꢀ17 cmÀ1). This is consistent with previ-
ously reported results concerning the protonation of adt
Scheme 1.
complexes, such as [Fe2(CO) CHTUNGTRNE{NNUG (m-SCH2)2NCH2C6H4X}] (X=
6A
2-Br, 4-Br) and [Fe2(CO)
6ACTHNUGTREN{NUNG (m-SCH2)2NCH2R}] (R=phenyl
tion of 6 yields the FeIFeII hydride 11. Reprotonation of the
nitrogen then gives complex 12, in which the interaction be-
tween the N-proton and the hydride makes the system pre-
disposed for hydrogen evolution. In principle, this could
occur directly, but it is more likely facilitated by a second
electron transfer to give the FeIFeI hydride 13. Again, this
electron transfer could be proton-coupled, facilitating the
release of molecular hydrogen and giving 13 only as a short-
lived intermediate. During or after these events, solvation of
the formed hydroxide ions and their replacement by water
regenerates the starting complex 6.
or alkyl).[6]
For complex 6, addition of one equivalent of HOTf result-
ed in a shift of the n(CO)Fe bands to become a broad feature
at about 2030 cmÀ1. This large shift (ca. 50 cmÀ1) is charac-
À
teristic of protonation at the Fe Fe bond (Figure 2b). On
addition of five equivalents of HOTf, two well-defined
n(CO)Fe bands were seen at 2032 and 1993 cmÀ1, clearly
À
showing protonation of the Fe Fe bond (Figure 2c). Upon
further addition of HOTf up to 30 equivalents, these bands
were shifted to 2046 and 2007 cmÀ1. This additional shift,
Dn(CO)Fe =14 cmÀ1, suggests formation of the hydride
Chem. Eur. J. 2010, 16, 2537 – 2546
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2539

B. ꢀkermark et al.
Figure 1. ORTEP diagrams (ellipsoids at 50% probability) of complexes
1 (a), 2 (b), and 3 (c).
Figure 2. IR spectra of 5 (left) and 6 (right): a) in CH3CN, b) +1 equiv of
HOTf, c) + 5 equiv of HOTf, d) + 30 equiv of HOTf.
[6(NH)ACHTUNGTRNEUNG
(FeHFe)]2+ (Figure 2d). The frequencies, and the
shifts thereof, for the two n(CO)Fe bands are consistent with
those reported for the diprotonated species of the related
complexes [Fe2(CO)
4ACHUTNGTREN(NUG PMe3)2ACHTNUGTRENNUNG
{(m-SCH2)2NCH2C6H5}] (9) and
[Fe2(CO)4A(PMe3)2A{(m-SCH2)2NCH2C6H4-2-Br}]
C
CHTUNGTRENNUNG
(10).[6a,8]
Upon addition of Et3N, the n(CO)Fe bands at 2046 and
2007 cmÀ1 reverted to 2032 and 1993 cmÀ1, showing clean
formation of the hydride [6ACHTNUGRTNEUNG
(FeHFe)]+.
Following the addition of 20 equivalents of HOTf to a so-
lution of 6 in CD3CN, the two protonated species [6-
ACHUTNGRENNUG CAHTUNGTRENNUGN
(FeHFe)]+ and [6(NH)(FeHFe)]2+ were present simultane-
ously according to in situ 1H and 31P NMR spectrometric
analyses. The hydride triplet signal at d=À15.59 ppm (cou-
pling constant JPH =21.4 Hz) is ascribed to [6ACTHNUTRGNEUNG
(FeHFe)]+, and
the second hydride signal at d=À15.57 ppm (coupling con-
stant JPH =22.3 Hz) is ascribed to [6(NH)
AHCTUNGTRENNUNG
values are very close to those reported previously for some
related m-H complexes.[2a,6a] The 31P NMR signals appeared
at d=21.5 ppm for [6
(FeHFe)]+ and at d=26.3 and
ACHTUNGTRENNUNG
22.3 ppm for [6(NH)
AHCTUNGTRENNUNG
spectra are thus consistent with the conclusions drawn from
the IR spectra.
Complex 5 behaved in a similar way as complex 6
(Figure 3). The suggested protonation process is reasonable,
since the pKa of the N-protonated species is around 10–12,
while the pKa of the Fe hydride is around 15.[2a,3a] For com-
plex 4, the protonation process was slightly different and
similar to that described for complexes 9 and 10.[6a,8] As with
these, the N-protonated species was the initial, metastable
product, which spontaneously rearranged to the thermody-
namically more stable Fe hydride species. This suggests that
the carboxyl group does not influence the protonation of
À
complex 4. In contrast, the direct protonation of the Fe Fe
bond in 5 and 6 could be due to the fact that the nitrogen is
engaged in hydrogen bonding to the carboxylate groups.
2540
www.chemeurj.org
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 2537 – 2546

ACHTUNGTRENNUNG[FeFe]-Hydrogenase Model Complexes with Carboxylate Side-Chains
FULL PAPER
had merged with the peak at about À1.5 V. Since this peak
was only observed for complex 2, it could perhaps be related
to participation of the carboxyl group. Previously published
structures of such N-protonated complexes with a nitrogen
substituent suggest that upon protonation and subsequent
sp3 hybridization at the nitrogen, the proton will interact
with the iron cluster and the substituent will occupy an axial
position, pointing away from the cluster. However, a crude
molecular model suggests that in the case of 2 the carbon
chain is sufficiently long for interaction between the carbox-
yl group and the iron cluster even in this conformation. In
principle, therefore, the peak could be the result of proton-
coupled one-electron reduction of the complex, assisted by
the carboxyl group.
Although the current increased only marginally with in-
creasing acidity, on-line MS analysis of controlled-potential
electrolysis at E=À1.25 V in a closed system in the pres-
ence of HOTf showed that complexes 1–3 catalyzed the for-
mation of hydrogen at this potential. This suggests that both
Figure 3. Hydride signals in
(FeHFe)]+ and [6(NH)(FeHFe)]2+ (obtained after treatment of complex
6 with 20 equivalents of HOTf).
a
1H NMR spectrum attributable to [6-
G
ACHTUNGTRENNUNG
Repeated attempts were made to obtain single crystals of
the
initial
monoprotonated
complexes,
such
as
[6
(FeHFe)]+ and [6(NH)
G
[(FeIFeI)(NH)]+, and the intermediate [(HFeIIFeI)(NH)]+
are reduced at around À1.25 V. Similar results have recently
been obtained by Talarmin, Gloaguen, and co-workers,
using a complex with a methoxyethyl substituent on the ni-
trogen in place of the carboxylate groups presented here.[3b]
FTIR and CV analyses of the solid isolated from the solu-
tion after the electrolysis showed that the unchanged start-
ing material could be recovered. The complexes 1–3 thus
clearly act as catalysts and a CECE (chemical–electrochemi-
cal–chemical–electrochemical) mechanism (A) seems rea-
sonable, in accordance with a recent proposal.[3b] However,
the observed redox potentials in the presence of strong acid
also indicate the involvement of a second mechanism (B)[3b]
(see the Supporting Information).
attempts failed, partly because the diprotonated species in
particular was unstable in solution above 263 K. Solutions of
the diprotonated forms of 6 and 5 spontaneously generated
the m-hydride complexes, for example [6ACTHNUTRGNEUNG
(FeHFe)]+, together
with decomposition products, the structures of which have
not been established.
Electrochemistry of complexes 1–3 under strongly acidic
conditions (HOTf): Complexes 1–3 each displayed a reduc-
tion peak at about À1.6 V versus Fc/Fc+ (Figure 4, Table 3).
Table 3. Electrochemical data of complexes 1–6 (V vs. Fc/Fc+).
ox
Complex
Epa FeIFeI/FeIIFeI
Epc FeIFeI/Fe0FeI
1
2
3
4
5
6
+0.62
+0.62
+0.64
À0.23
À0.27
À0.12
À1.61
À1.60
À1.63
À2.01
À2.05
À1.94
Electrochemistry of complexes 4–6 under strongly acidic
conditions (HOTf): Complexes 4–6, in which two CO li-
gands have been replaced by PMe3, were also studied by
CV in the presence of HOTf. Before the addition of acid,
complexes 4 and 5 were reduced at potentials close to À2 V,
while 6 was reduced at À1.9 V. Related complexes lacking
the carboxyl group are generally reduced at potentials
around À2.2 V.[2a,8,10] It thus appears that the carboxyl
groups contribute to lowering the potential for electron
transfer, especially in complex 6, in which the interaction
should be strongest for steric reasons, in accordance with
the relevant carboxyl IR frequencies (Table 1).
On addition of one or two equivalents of triflic acid, their
reduction peaks moved to about À1.20 V. This potential,
about 400 mV more positive than the reduction potential of
the starting materials, can be explained in terms of protona-
tion of the nitrogen bridge.[2] New reduction waves appeared
at À1.54 V for 1 and 2, and at À1.56 V for 3. Upon the addi-
tion of more acid (3–10 equiv), the peak at À1.20 V was es-
sentially unaffected, but for the new waves the currents
grew linearly with increasing amount of acid. The peaks also
shifted slightly to about À1.5 V. Similar results have recently
Upon addition of acid, the expected peak shifts were ob-
served. For 4, two reduction peaks were recorded, at À1.46
and À1.90 V. The shift of the peak at À1.46 V is of the same
order (440 mV) as that observed upon protonation at nitro-
gen in related complexes such as [Fe2(CO)4ACHTNUTRGENN(UG PMe3)2ACHTUNGTRENNUNG{(m-
been reported for the azadithiolate complex [Fe2(CO)
{m-
SCH2)2NCH2C6H4-2-Br}].[6a] Neither the magnitude nor the
position of this peak changed significantly on increasing the
acid concentration. The peak at À1.90 V also retained its po-
sition, but the current was directly proportional to the
amount of added HOTf. This indicates that the N-protonat-
SCH2N(CH2CH2OMe)CH2S}].[2b] For complex 2, a third,
sharp peak appeared at À1.36 V after addition of three
equivalents of acid. This peak shifted with increasing acid
concentration and after the addition of five equivalents it
Chem. Eur. J. 2010, 16, 2537 – 2546
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2541

B. ꢀkermark et al.
Figure 4. Cyclic voltammetry of complexes 1–6 (1 mm) in 0.05m nBu4NPF6/CH3CN in the absence and presence of triflic acid at a scan rate of
100 mVsÀ1
.
ed species is electrochemically active in the reduction of
protons. All of the reduction currents displayed a linear de-
pendence of ip on v1/2, suggesting one-electron processes.
Complexes 5 and 6 behaved slightly differently. Only after
addition of two equivalents of HOTf did new sharp peaks
appear at À1.52 V for 6 (anodic shift 420 mV) and at
À1.50 V for 5 (shift 550 mV). The intensities of these peaks
increased only slightly on further addition of acid, but their
potentials moved to more negative values, while at the same
time a broad underlying reduction peak rapidly increased,
starting at about À1 V. The increase of current vs amount of
added acid was considerably greater for 5 and 6 (ca. 200 mA
after addition of 4 equiv of TfOH) than for 4 (ca. 50 mA).
For both 5 and 6, curve crossing was observed after addition
of about five equivalents of acid. The implication of this is
not clear, but it suggests that a chemical reaction, such as re-
ductive elimination of hydrogen, gives an intermediate that
is oxidized at the crossing potential, around À1.1 V. The
electrochemical activity at each of the reduction peaks dis-
played a first-order dependence on [HOTf], and a one-elec-
tron reduction process was also indicated by the linear de-
pendence of ip on v1/2. A simple CECE mechanism seems
reasonable for the formation of hydrogen with these com-
plexes (Scheme 2).
Light-driven hydrogen production: Light-driven hydrogen
generation was studied in a similar way as previously report-
ed, with a three-component system of [RuACTHNUTRGNEUNG
(bpy)3]2+, a diiron
complex, and ascorbic acid (H2A) in CH3CN/D2O solu-
tion.[11] On-line MS was used to monitor the evolved gases
in the molecular weight range 0–100. Hydrogen also ap-
peared to be formed through processes other than the re-
2542
www.chemeurj.org
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 2537 – 2546

ACHTUNGTRENNUNG[FeFe]-Hydrogenase Model Complexes with Carboxylate Side-Chains
FULL PAPER
Figure 6. Isotope distribution of the evolved hydrogen, studied by on-line
Scheme 2.
gas analysis by recording mass spectra. ([RuACTHNUTRGNEUNG
(bpy)3]2+, 1 mmol; complex 1,
5 mmol; H2A, 1 mmol; CH3CN/D2O, 3.75 mL:1.25 mL).
duction of protons from water, for example from acetoni-
trile in the mass spectrometer. Deuterated water was there-
fore used in order to ensure that the hydrogen produced
really derived from the added water.
The MS analysis of total evolved hydrogen (H2, HD, and
D2) as a function of reaction time for complexes 1–3 is
shown in Figure 5. The initial slope of the hydrogen genera-
20 min. This could explain the low TN number for hydrogen
evolution, since it indicates photo-induced decomposition of
the Fe2S2 complexes. Complexes 2 and 3 displayed similar
photo-catalytic properties (Figure 6 and the Supporting In-
formation).
Electrochemistry in aqueous solution: Finally, since Nature
obviously cannot make use of strong acids, we were interest-
ed in ascertaining whether the carboxyl groups could influ-
ence the redox chemistry of the complexes 2, 3, 5, and 6 in
aqueous solution. With complex 2, the effect appeared to be
small, since the reduction peak at À1.6 V was essentially un-
affected by addition of water to the solution in acetonitrile
generally used. The second peak at around À2.2 V was also
only marginally changed (Figure 7). The reduction potential
of complex 3 at À1.6 V behaved in a similar way as that of
complex 2, although a small, steady increase in the current
could be observed upon addition of increasing amounts of
water. In contrast, at the second reduction of complex 3,
which was found at a less negative potential than in the case
of 2 (À1.9 V), a sharp increase in the current was observed
upon addition of 300 or 400 mL of water. Although in the
crystal structure 3 adopts a conformation in which the car-
boxyl group points away from the nitrogen, rotation to give
a conformation with hydrogen bonding to the nitrogen
could conceivably take place in a polar solvent, which would
break the dimeric structure. Because the nitrogen here is
not a strong hydrogen-bond acceptor, only a fairly small
effect on redox behavior might be expected, in accordance
with the experimental findings.
With complex 5, a sharp increase in current was noted at
about À2.3 V on addition of the first 50 mL of water, but
subsequent addition had little further effect. Complex 6 be-
haved differently, however. A small peak was observed at
around À1.7 V, which increased slightly and moved to about
À1.5 V with increasing addition of water. However, the
most prominent effect was a sharp increase in current, start-
ing at around À2 V, as more water was added. Decreases in
the negative redox potentials for both 3 and 6 by about
300 mV to around À2 V, accompanied by sharp increases in
the currents at these potentials, were observed (Figure 7).
Figure 5. Evolved hydrogen studied by on-line gas analysis by recording
mass spectra. ([Ru
CH3CN/D2O, 2.5 mL:2.5 mL).
ACHTUNGTRENNUNG , 1 mmol; Fe2S2, 5 mmol; H2A, 1 mmol;
(bpy)3]2+
tion rate for complex 2 is about twice that for complex 3
and four times that for complex 1. However, after about
40 min, almost the same amount of hydrogen had been
formed by complexes 2 and 3, close to three times that pro-
duced with 1.
Analysis of hydrogen formation (H2, HD, and D2) as a
function of reaction time for complex 1 showed that the
amount of hydrogen evolved first increased steadily and
then leveled off after around 90 min at about 1 mmol
(Figure 6). The total amount of D2 evolved was about twice
that of HD and thrice that of H2. H2 formation was essen-
tially complete after 20 min, while the evolution of HD and
D2 continued for another 70 min. With complex 1, the initial
slopes for hydrogen evolution were about 23 nmolminÀ1 for
H2, 20 nmolminÀ1 for HD, and 45 nmolminÀ1 for D2.
Another important feature is that carbon monoxide was
detected with a turnover number (TN) close to 3 after
Chem. Eur. J. 2010, 16, 2537 – 2546
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2543

B. ꢀkermark et al.
Figure 7. Cyclic voltammetry of complexes 2, 3, 5, and 6 (1 mm) in 0.05m nBu4NPF6/CH3CN in the presence of water at a scan rate of 100 mVsÀ1
.
These results are similar to those obtained by Lichtenberger
and Darensbourg and their co-workers following the addi-
tion of acetic acid to Fe2S2 complexes.[10] This shows that, at
least in complex 6, the carboxyl group does indeed play a
role in the reduction of protons to molecular hydrogen.
through proton transfer to the Fe2S2 site.[1e,12,13] They might
also be important for regulation of the potential for proton
reduction, as suggested by the results for complexes 5 and 6.
Although this effect is small, it opens a fresh perspective in
the construction of new types of hydrogenase models.
Conclusion
Experimental Section
In conclusion, a series of mimics of the active site of [FeFe]-
hydrogenase, specifically adt-bridged Fe2S2 complexes 1–6
bearing carboxylic acid functionalities, has been prepared.
Photoinduced hydrogen evolution is promoted by these
complexes, and analysis by means of on-line MS showed the
formation of D2, HD, and H2, although the turnover number
was still low.
On treatment with triflic acid, complexes 1–3 are proton-
ated exclusively at the nitrogen. In contrast, complexes 4–6,
in which two carbonyl ligands have been substituted by
PMe3 groups, undergo different protonation processes. In 4,
General remarks: Unless otherwise indicated, all reactions and opera-
tions related to organometallic complexes were carried out under a dry,
oxygen-free argon atmosphere using standard Schlenk techniques. Chem-
icals were purchased from Aldrich, and all solvents were dried and dis-
tilled prior to use according to the standard methods.
Infrared spectra were recorded on a Perkin–Elmer spectrum 1 spectrom-
eter. NMR spectra were collected on a Varian spectrometer (400 MHz).
Elemental analyses were performed at the Analytische Laboratorien,
Lindar, Germany. Electro-spray mass spectrometry (FTICR-ESI-MS)
was performed on a Bruker Daltonics BioAPEX-94e superconducting
9.4 Tesla FTICR mass spectrometer (Bruker Daltonics, Billerica, MA,
USA).
Acetonitrile (Aldrich, spectroscopy grade) used in the electrochemistry
experiments was dried with molecular sieves (4 ꢀ) and then freshly dis-
tilled from CaH2 under an argon atmosphere. A solution of 0.05m
nBu4NPF6 (Fluka, electrochemical grade) in CH3CN was used as electro-
lyte. Electrochemical measurements were carried out in a three-electrode
cell connected to an Autolab potentiostat with a GPES electrochemical
interface. The working electrode was a glassy carbon disc (diameter
2 mm, freshly polished). The reference electrode was a non-aqueous Ag/
Ag+ electrode (1.0 mm AgNO3 in CH3CN) with the ferrocene/ferroceni-
um (Fc/Fc+) couple as an external standard. The auxiliary electrode was
a platinum wire. All potentials reported herein are quoted relative to the
À
protonation occurs first on the bridging N, whereas the Fe
Fe bond is protonated first in 5 and 6. Upon addition of
strong acid, each of the complexes catalyzed the reduction
of protons to hydrogen.
Interestingly, complexes 5 and 6 proved to be electro-
chemically active upon the addition of water alone, suggest-
ing that the carboxyl group can promote proton reduction
under neutral conditions. In natural hydrogenases, a number
of neighboring amino acids are believed to participate
2544
www.chemeurj.org
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 2537 – 2546

ACHTUNGTRENNUNG[FeFe]-Hydrogenase Model Complexes with Carboxylate Side-Chains
FULL PAPER
Fc/Fc+ couple. The solutions were purged with solvent-saturated argon
to remove residual oxygen. All experimental measurements were made
while maintaining the systems under an atmosphere of argon at all times.
these components in the enclosed volume. This permitted a quantitative
determination of the amounts of gases produced with masses in the inter-
val 1–100 mass units.
In studies of photochemical hydrogen generation, a glass tube was at-
tached to the enclosed volume (Figure 8). Deuterated water/acetonitrile
(5 mL), [RuACHTUNGTRNNEUG(bpy)3ACHTUNGTERNN(UGN PF6)2] (1 mmol), H2A (1 mmol), and complex 1, 2, or 3
Single-crystal X-ray diffraction patterns were recorded with an Oxford
Diffraction Excalibur diffractometer equipped with a sapphire-3 CCD
using an MoKa radiation source (l=0.71073 ꢀ) with w scans at different
f values to fill the Ewald sphere. The sample-to-detector distance was
50 mm. The maximum value of 2q was 638.
(5 mmol) were mixed in a test tube. The total volume of water and aceto-
nitrile was kept at 5 mL, but with ratios of water to acetonitrile of 1:30,
1:3, 1:1, and 3:1. The test tube was then attached to the enclosed volume,
which was evacuated for ~5 min to minimize the amount of dissolved
gases in the solution. Ar at ~30 mbar was then introduced into the en-
closed volume. The inlet to the MS was adjusted, and after a period of at
least 10 min, during which the partial pressures in the MS stabilized, the
light was switched on. In this way, the start of the experiment was well
defined. The light source used was a 150 W halogen lamp. As the reac-
tions were conducted in a glass vessel, the mixtures could be illuminated
directly through the glass walls of the reactor. In some experiments, a
Pyrex glass UV filter was used, but this had negligible effect on the reac-
tions.
Indexing, cell refinements, and integration of reflection intensities were
performed with Crysalis software.[14] Numerical absorption correction
was performed with the program X-RED,[15] verifying the crystal shape
with the program X-shape.[16] The structure was solved by direct methods
using SHELXS-97,[17] which yielded electron density maps in which most
of the non-hydrogen atoms could be resolved. The rest of the non-hydro-
gen atoms were located from difference electron density maps, and the
structure model was refined through full-matrix least-squares calculations
on F2 using the program SHELXL97–2.[18] All non-hydrogen atoms were
refined with anisotropic displacement parameters, while the hydrogens,
which were placed in geometrically calculated positions and allowed to
ride on the atoms to which they were bonded, were given isotropic dis-
placement parameters calculated as x·(Ueq for the non-hydrogen atoms),
with x=1.2 for methylenic (-CH2-) and aromatic hydrogens.
H2NCH2C6H4-2-COOH·HCl: The starting compound phthalimidine
(1.33 g, 10 mmol), prepared according to a literature procedure,[24] was re-
fluxed in 37% HCl (140 mL) for 10 h. Removing the solvent under re-
duced pressure and washing the residue with CHCl3 (20 mL) afforded
the target (2-benzylamino)carboxylic acid (HCl salt) as a white solid
(1.31 g, 70%). 1H NMR (D2O): d=4.27 (s, 2H; CH2), 7.35–7.50 (m, 2H;
C6H4), 7.57 (t, J=7.4 Hz, 1H; C6H4), 7.97 ppm (d, J=7.6 Hz, 1H; C6H4);
MS (ESI-TOF): m/z: 152.072 [M+H]+.
CCDC-744052 (1), CCDC-744053 (2), and CCDC-744054 (3) contain the
supplementary crystallographic data for this paper. These data can be ob-
tained free of charge from The Cambridge Crystallographic Data Centre
via www.ccdc.cam.ac.uk/data_request/cif.
Gas analysis by mass spectrometry: The mass spectrometer set-up used
consisted of three separate parts connected with gas valves[19–23] (see
Figure 8), namely an enclosed volume (or reaction chamber), a gas han-
ACHUTNGRENUN[G Fe2(CO)6ACHTUTGNERN{NNGU (m-SCH2)2NCH2C6H4-2-COOH}] (1): A solution of (2-benzyla-
mino)carboxylic acid (HCl salt) (0.25 g, 1.3 mmol) and Et3N (0.18 mL,
1.3 mmol) in THF (40 mL) was stirred for 0.5 h at room temperature and
then cooled to 08C. A solution of [Fe2(CO)6ACHTNUGRTNEUNG
(HOCH2S)2][5] (1.28 mmol)
in THF (20 mL) was then added. After allowing the reaction to proceed
for 5 h, the solution was filtered and the solvent was removed on a rotary
evaporator. The crude product was purified by column chromatography
with CH2Cl2/CH3COCH3 (10:1, v/v) as eluent to give 1 (200 mg, 30%) as
a red solid. 1H NMR (CDCl3): d=3.68 (s, 4H; 2ꢃNCH2S), 4.35 (s, 2H;
C6H4CH2N), 7.32 (d, J=7.5 Hz, 1H; C6H4), 7.42 (t, J=7.5 Hz, 1H;
C6H4), 7.57 (t, J=7.4 Hz, 1H; C6H4), 8.11 ppm (d, J=7.6 Hz, 1H; C6H4);
IR (CH2Cl2): n(CO)=2073, 2029, 1991, 1691 cmÀ1; elemental analysis
calcd (%) for C16H11Fe2NO8S2: C 36.88, H 2.13, N 2.69; found: C 37.04,
H 2.43, N 2.95.
A
{(m-SCH2)2NCH2CH2CH2COOH}] (2): 4-Aminobutyric acid
solution of [Fe2(CO)6-
a
AHCTUNGTRENNUNG
room temperature, the solution was filtered and then the solvent was re-
moved on a rotary evaporator. The crude product was purified by
column chromatography with CH2Cl2/CH3COCH3 (10:1, v/v) as eluent to
give 2 (389 mg, 41%) as a red solid. 1H NMR (CD3CN): d=1.70 (s, 2H;
CH2CH2CH2), 2.39 (m; CH2COOH), 2.79 (s, 2H; CH2CH2N), 3.56 ppm
(s, 4H; NCH2S); IR (CH2Cl2): n(CO)=2073, 2029, 1991, 1734 cmÀ1; ele-
mental analysis calcd (%) for C12H11Fe2NO8S2: C 30.47, H 2.34, N 2.96;
found: C 30.81, H 2.75, N 3.14.
Figure 8. The experimental set-up for gas analysis by recording mass
spectra.
dling system (GHS), and a mass spectrometer (MKS Spectra Products,
Microvision Plus, 0–100 mass units) under ultra-high vacuum (base pres-
sure 2ꢃ10À10 bar). The GHS permitted control of the atmosphere in the
enclosed volume. Almost any gas reservoir could be connected to the
GHS and, in principle, the only limitation was that the vapor pressure of
the gases used had to be higher than 1 mbar. A rough pump was used to
evacuate the GHS, so that the pressure could be regulated in the range
0.1–1000 mbar. With this set-up, the enclosed volume was continuously
probed by the mass spectrometer through an inlet regulated by a leak
valve. The inlet to the mass spectrometer was so small that the probing
caused a negligible pressure change in the enclosed volume. Consequent-
ly, the measurements had a negligible influence on the measured reaction
rates. The measured data comprised partial pressures in the MS of
masses 0–100 versus time. By calibration of the system, these measured
data could be converted into partial pressures of the different compo-
nents in the enclosed volume. As the volume and temperature of the en-
closed volume were known, by applying the gas laws the partial pressures
of the different components could be converted into the amounts of
A
6ACHTUNGTERNNUG{(m-SCH2)2NCHACTHUNGTRENNNUG
a
A
room temperature, the solution was filtered and then the solvent was re-
moved on a rotary evaporator. The crude product was purified by
column chromatography with CH2Cl2/CH3COCH3 (10:1, v/v) as eluent to
give 3 (389 mg, 41%) as a red solid. 1H NMR (CD3CN): d=0.85 (t, J=
7.5 Hz, 3H; CH3), 1.71 (m, 2H; CH3CH2CH), 3.16 (t, J=7.3 Hz, 1H;
CH2CHNCOOH), 3.37 (d, J=12.0 Hz, 2H; NCH2S), 3.47 ppm (d, J=
12.0 Hz, 2H; NCH2S); IR (CH2Cl2): n(CO)=2074, 2033, 1994, 1710 cmÀ1
elemental analysis calcd (%) for C12H11Fe2NO8S2: C 30.47, H 2.34, N
2.96; found: C 30.91, H 2.69, N 3.15.
ACHNUTGERTG[NUNN Fe2(CO)4AHCTNUTREGG(NNNU PMe3)2CAHTGUNRTEN{NUGN (m-SCH2)2NCH2C6H4-2-COOH}] (4): A solution of 1
(0.052 g, 0.1 mmol) and PMe3 (0.4 mL, 0.4 mmol, 1m in THF) in THF
(20 mL) was refluxed for 3.5 h. The solvent was then removed and the
crude product was purified by column chromatography with CH3COCH3/
Chem. Eur. J. 2010, 16, 2537 – 2546
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2545

B. ꢀkermark et al.
CH3OH/Et3N (200:10:1, v/v/v) as eluent to give
4
(32 mg, 51%).
[4] a) C. M. Thomas, O. Rꢆdiger, T. Liu, C. E. Carson, M. B. Hall, M. Y.
Darensbourg, Organometallics 2007, 26, 3976–3984; b) V. Vijai-
kanth, J. F. Capon, F. Gloaguen, F. Y. Pꢅtillon, P. Schollhammer, J.
Talarmin, J. Organomet. Chem. 2007, 692, 4177–4181.
[5] a) H. Li, T. B. Rauchfuss, J. Am. Chem. Soc. 2002, 124, 726–727;
b) J. L. Stanley, Z. M. Heiden, T. B. Rauchfuss, S. R. Wilson, L. De
Gioia, G. Zampella, Organometallics 2008, 27, 119–125.
[6] a) F. Wang, M. Wang, X. Liu, K. Jin, W. Dong, L. Sun, Dalton Trans.
2007, 3812–3819; b) S. Ott, M. Kritikos, B. ꢀkermark, L. Sun, R.
Lomoth, Angew. Chem. 2004, 116, 1024–1027; Angew. Chem. Int.
Ed. 2004, 43, 1006–1009; c) W. Gao, J. Liu, C. Ma, L. Weng, K. Jin,
C. Chen, B. ꢀkermark, L. Sun, Inorg. Chim. Acta 2006, 359, 1071–
1080; d) W. Gao, J. Ekstrçm, J. Liu, C. Chen, L. Eriksson, L. Weng,
B. ꢀkermark, L. Sun, Inorg. Chem. 2007, 46, 1981–1991.
[7] B. ꢀkermark, Acta Chem. Scand. 1961, 15, 985–990.
[8] L. Schwartz, G. Eilers, L. Eriksson, A. Gogoll, R. Lomoth, S. Ott,
Chem. Commun. 2006, 520–522.
[9] a) W. Gao, J. Liu, B. ꢀkermark, L. Sun, Inorg. Chem. 2006, 45,
9169–9171; b) L. Song, J. Ge, X. Liu, L. Zhao, Q. Hu, J. Organomet.
Chem. 2006, 691, 5701–5709; c) L. Song, H. Wang, J. Ge, S. Mei, J.
Gao, L. Wang, B. Gai, L. Zhao, J. Yan, Y. Wang, Organometallics
2008, 27, 1409–1416; d) K. Justice, G. Zampella, L. De Gioia, T. B.
Rauchfuss, J. I. Van der Vlugt, S. R. Wilson, Inorg. Chem. 2007, 46,
1655–1664.
[10] a) D. Chong, I. P. Georgakaki, R. Meija-Rodriguez, J. Sannabria,
M. P. Soriaga, M. Y. Darensbourg, Dalton Trans. 2003, 4158–4163;
b) G. A. N. Felton, A. K. Vanucci, J. Chen, L. T. Lockett, N. Oku-
mura, B. J. Petro, U. I. Zakai, D. H. Evans, R. S. Glass, D. L. Lich-
tenberger, J. Am. Chem. Soc. 2007, 129, 12521–12530.
[11] a) Y. Na, M. Wang, J. Pan, P. Zhang, B. ꢀkermark, L. Sun, Inorg.
Chem. 2008, 47, 2805–2810; b) X. Li, M. Wang, S. Zhang, J. Pan, Y.
Na, J. Liu, B. ꢀkermark, L. Sun, J. Phys. Chem. B 2008, 112, 8198–
8202.
[12] Y. Nicolet, C. Piras, P. Legrand, C. E. Hatchikian, J. C. Fontecilla-
Camps, Structure 1999, 7, 13–23.
[13] J. W. Peters, W. N. Lanzilotta, B. J. Lemon, L. C. Seefeldt, Science
1998, 282, 1853–1858.
1H NMR (CD3CN): d=1.50 (d, J=9.2 Hz, 18H; PCH3), 3.21 (s, 4H; 2ꢃ
NCH2S), 4.09 (s, 2H; C6H4CH2N), 7.28 (s, 3H; C6H4), 7.63 ppm (s, 1H;
C6H4); 31P NMR (CD3CN): d=22.33 ppm; IR (CH2Cl2): n(CO)=1977,
1939, 1899, 1694 cmÀ1
;
elemental analysis calcd (%) for
C20H31NFe2O6P2S2: C 38.79, H 5.05, N 2.26; found: C 39.25, H 5.51, N
1.79.
ACHTUNGTRENNUNG[Fe2(CO)4ACHTUNGTRENNUNG(PMe3)2ACHTUNGTRENNUNG{(m-SCH2)2NCH2CH2CH2COOH}] (5): A solution of 2
(0.047 g, 0.1 mmol) and PMe3 (0.4 mL, 0.4 mmol, 1m in THF) in THF
(20 mL) was refluxed for 3.5 h. The solvent was then removed and the
crude product was purified by column chromatography with CH3COCH3/
CH3OH/Et3N (200:10:1, v/v/v) as eluent to give
5 (29 mg, 51%).
1H NMR (CD3CN): d=1.53 (d, J=9.3 Hz, 18H; PCH3), 1.97 (m, 2H;
CH2CH2CH2), 2.36 (s, 2H; CH2COOH), 2.51 (s, 2H; CH2CH2N),
3.23 ppm (s, 4H; 2ꢃNCH2S); 31P NMR (CD3CN): d=22.36 ppm; IR
(CH2Cl2): n(CO)=1979, 1938, 1896, 1665 cmÀ1; elemental analysis calcd
(%) for C16H29Fe2NO6P2S2: C 33.76, H 5.14, N 2.46; found: C 34.05, H
5.53, N 2.49.
ACHTUNGTRENNUNG[Fe2(CO)4ACHTUNGTRENNUNG(PMe3)2ACHTUNGTRENNUNG{(m-SCH2)2NCHACHTNUGTRNEN(UGN CH2CH3)COOH}] (6): A solution of 3
(0.047 g, 0.1 mmol) and PMe3 (0.4 mL, 0.4 mmol, 1m in THF) in THF
(20 mL) was refluxed for 3.5 h. The solvent was then removed and the
crude product was purified by column chromatography with CH3COCH3/
CH3OH/Et3N (200:10:1, v/v/v) as eluent to give
6 (28 mg, 50%).
1H NMR (CD3CN): d=1.32 (t, J=4.5 Hz, 3H; CH2CH3), 1.50 (s, 18H;
PCH3), 1.55 (m, 2H; CHCH2CH3), 3.41 (s, 1H; CH2CHNCOOH),
3.58 ppm (s, 4H; 2ꢃNCH2S); 31P NMR (CD3CN): d=22.7 ppm; IR
(CH2Cl2): n(CO)=1978, 1941, 1898, 1659 cmÀ1; elemental analysis calcd
(%) for C16H29Fe2NO6P2S2: C 33.76, H 5.14, N 2.46; found: C 34.05, H
5.53, N 2.49.
Protonation of complex 6: Complex 6 (5.7 mg, 0.01 mmol) was dissolved
in CD3CN (0.5 mL) in an NMR tube under argon atmosphere to give a
red solution. Incremental amounts of triflic acid (1 equiv: 0.88 mL) were
1
then added directly to the solution for in situ H NMR, 31P NMR, and IR
analyses. All attempts to isolate the protonated complexes were unsuc-
cessful.
[14] Xcalibur CCD system, CrysAlis Software system, Version 1.170,
Oxford Diffraction Ltd., Oxford, 2003.
[15] X-RED, absorption correction program, Version 1.09, Stoe & Cie
GmbH Darmstadt, Darmstadt.
Acknowledgements
[16] X-SHAPE, Crystal Optimisation for Numerical Absorption Correc-
tion, Version 1.2, Stoe & Cie GmbH Darmstadt, Darmstadt.
[17] G. M. Sheldrick, Acta Crystallogr. Sect. A 1990, 46, 467.
[18] SHELXL97–2, Computer program for the refinement of crystal
structures, G. M. Sheldrick, University of Gçttingen, Gçttingen,
1997.
[19] G. Hultquist, L. Grꢇsjç, Q. Lu, T. ꢀkermark, Corr. Sci. 1994, 36,
1459–1471.
[20] T. ꢀkermark, J.-J. Ganem, I. Trimaille, S. Rigo, J. Electrochem. Soc.
1999, 146, 4580–4585.
We are grateful to the Swedish Research Council, the K&A Wallenberg
Foundation, and the Swedish Energy Agency for financial support of this
work.
[1] a) A. L. De Lacey, V. M. Fernꢄndez, M. Rouset, R. Cammack,
Chem. Rev. 2007, 107, 4304–4330; b) P. M. Vignais, B. Billoud,
Chem. Rev. 2007, 107, 4206–4272; c) G. J. Kubas, Chem. Rev. 2007,
107, 4152–4205; d) W. Lubitz, E. Rejerse, M. Van Gastel, Chem.
Rev. 2007, 107, 4331–4365; e) J. C. Fontecilla-Camps, A. Volbeda, C.
Cavazza, Y. Nicolet, Chem. Rev. 2007, 107, 4273–4303; f) F. Gloa-
guen, T. B. Rauchfuss, Chem. Soc. Rev. 2009, 38, 100–108; g) C.
Tard, C. J. Pickett, Chem. Rev. 2009, 109, 2245–2274.
[2] a) G. Eilers, L. Schwartz, M. Stein, G. Zampella, L. De Gioia, S.
Ott, R. Lomoth, Chem. Eur. J. 2008, 14, 7075–7084; b) J.-F. Capon,
S. Ezzaher, F. Gloaguen, F. Y. Pꢅtillon, P. Schollhammer, J. Talarmin,
Chem. Eur. J. 2008, 14, 1954–1964; c) B. E. Barton, M. T. Olsen,
T. B. Rauchfuss, J. Am. Chem. Soc. 2008, 130, 16834–16835.
[21] T. ꢀkermark, B. Emmoth, H. Bergsꢇker, J. Nucl. Mater. 2006, 359,
220–226.
[22] C. Anghel, Q. Dong, J. Mater. Sci. 2007, 42, 3440–3453.
[23] Y. Xu, T. ꢀkermark, V. Gyollai, D. Zou, L. Eriksson, L. Duan, R.
Zhang, B. ꢀkermark, L. Sun, Inorg. Chem. 2009, 48, 2717–2719.
[24] K. Y. Koltunov, G. K. Surya Prakash, G. Rasul, G. A. Olah, Eur. J.
Org. Chem. 2006, 4861–4866.
[3] a) M. H. V. Huynh, T. J. Meyer, Chem. Rev. 2007, 107, 5004–5064;
b) C. Costentin, Chem. Rev. 2008, 108, 2145–2179.
Received: August 17, 2009
Published online: January 14, 2010
2546
www.chemeurj.org
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 2537 – 2546